Author name cluster

Kaushik Roy

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

34 papers

2 author rows

TMLR Journal 2026 Journal Article

Achieving Global Flatness in Decentralized Learning with Heterogeneous Data

Sakshi Choudhary
Sai Aparna Aketi
Kaushik Roy

Decentralized training enables peer-to-peer on-device learning without relying on a central server, but suffers from degraded generalization performance under heterogeneous data distributions due to local overfitting. One strategy to mitigate this is to seek flatter loss landscapes during local optimization at each client. However, with extreme data heterogeneity, local objectives may diverge from the global one, yielding local flatness rather than true global flatness. To mitigate this challenge, we introduce GFlat, a novel decentralized algorithm that enables each client to estimate and incorporate an approximation of the global update direction while seeking a flatter loss landscape locally. This lightweight strategy allows each client to directly contribute to global flatness without requiring additional communication or centralized coordination. We theoretically analyze the convergence properties of GFlat and validate its performance through extensive experiments across a range of datasets, model architectures, and communication topologies. GFlat consistently improves generalization in non-IID data settings and achieves up to 6.75\% higher test accuracy compared to state-of-the-art decentralized methods.

TMLR Journal 2026 Journal Article

From Clutter to Clarity: Visual Recognition through Foveated Object-Centric Learning (FocL)

Amitangshu Mukherjee
Deepak Ravikumar
Kaushik Roy

Human active vision integrates spatial attention (dorsal) and object recognition (ventral) as distinct information processing pathways. Rapid eye movements focus perception on task-relevant regions while filtering out background clutter. Mimicking this ventral specialization, we introduce FocL (Foveated Object-Centric Learning), a training strategy that biases image classification models toward label-consistent object regions by replacing full images with foveated crops. Standard training often relies on spurious correlation between label and background, increasing memorization of hard examples in the tail of the difficulty distribution. FocL simulates saccades by jittering fixation points and extracting foveated glimpses from annotated bounding boxes. This object-first restructuring reduces non-foreground contamination and lowers mean training loss. FocL reduces memorization, lowering mean cumulative sample loss by approximately 65 % and making nearly all high-memorization samples (top 1 %) easier to learn. It also increases the mean $\ell_2$ adversarial perturbation distance required to flip predictions by approximately 62 %. On ImageNet-V1, FocL achieves up to 11 % higher accuracy on oracle crops. When paired with the Segment Anything Model (SAM) as a dorsal proposal generator, FocL provides around an 7 % gain on ImageNet-V1 and up to 8 % under natural distribution shift (ImageNet-V2). Extending this setup to COCO, FocL improves cross-domain mAP by 3--4 points without any target-domain training. Finally, given object localization (bounding boxes), FocL reaches higher accuracy using roughly 56\% fewer training images, offering a simple path to more robust and efficient visual recognition.

AAAI Conference 2026 System Paper

In-Situ Eval: A Modular Framework for Custom and Real-Time RAG Benchmarking

Ritvik Garimella
Kaushik Roy
Chathurangi Shyalika
Amit Sheth

Retrieval-Augmented Generation (RAG) has become the standard approach for integrating domain knowledge into Large Language Models (LLMs). However, fair comparison of RAG pipelines remains difficult: data preparation is often ad hoc, subsampling methods are opaque, parameters vary across implementations, and evaluation is fragmented. We present In-Situ Eval, a unified and reproducible framework that operationalizes the full RAG pipeline with configurable subsampling strategies and both RAG-specific and generic evaluation metrics. The platform supports two execution modes: an offline Dataset mode for evaluating precomputed outputs, and a live Retrieval mode for benchmarking RAG variants with state-of-the-art LLMs. Users can flexibly select datasets, retrieval techniques, models, and metrics, enabling side-by-side comparisons, ablations, and targeted analyses. This holistic approach reduces computational costs, clarifies the impact of subsampling techniques, and provides actionable insights for real-world deployments. By facilitating transparent, customizable, and interactive benchmarking, In-Situ Eval empowers both researchers and practitioners to make informed decisions in adapting RAG pipelines to domain-specific needs.

PDF Details DOI

TMLR Journal 2026 Journal Article

Learning to Localize Leakage of Cryptographic Sensitive Variables

Jimmy Gammell
Anand Raghunathan
Abolfazl Hashemi
Kaushik Roy

While cryptographic algorithms such as the ubiquitous Advanced Encryption Standard (AES) are secure, *physical implementations* of these algorithms in hardware inevitably `leak' sensitive data such as cryptographic keys. A particularly insidious form of leakage arises from the fact that hardware consumes power and emits radiation in a manner that is statistically associated with the data it processes and the instructions it executes. Supervised deep learning has emerged as a state-of-the-art tool for carrying out *side-channel attacks*, which exploit this leakage by learning to map power/radiation measurements throughout encryption to the sensitive data operated on during that encryption. In this work we develop a principled deep learning framework for determining the relative leakage due to measurements recorded at different points in time, in order to inform *defense* against such attacks. This information is invaluable to cryptographic hardware designers for understanding *why* their hardware leaks and how they can mitigate it (e.g. by indicating the particular sections of code or electronic components which are responsible). Our framework is based on an adversarial game between a classifier trained to estimate the conditional distributions of sensitive data given subsets of measurements, and a budget-constrained noise distribution which probabilistically erases individual measurements to maximize the loss of this classifier. We demonstrate our method’s efficacy and ability to overcome limitations of prior work through extensive experimental comparison on 6 publicly-available power/EM trace datasets from AES, ECC and RSA implementations. Our PyTorch code is available at https://github.com/jimgammell/learning_to_localize_leakage.

NAI Journal 2026 Journal Article

Metatuning: An Empirical Study of Judge-Guided Prompt Refinement and Its Boundary Conditions

Aniruddha Chattopadhyay
Raj Dandekar
Kaushik Roy

Iterative prompt refinement is a practical approach for improving the reliability of large language models without weight updates. In this work, we study metatuning: a judge-guided prompt-refinement loop in which an evaluator critiques errors and provides targeted natural-language corrections or demonstrations that are incorporated into the prompt. We evaluate metatuning on axiomatic deductive reasoning (MATH-500), on combinations with chain-of-thought and self-reflection prompting, and on video-based physical reasoning (CLEVRER). Our results show that metatuning can improve baseline performance in static, rule-like domains, but offers limited benefit when paired with strong reasoning baselines and does not generalize to spatiotemporal video reasoning. Overall, we identify boundary conditions for judge-guided prompt refinement and motivate future work on integrating feedback at the level of reasoning traces.

TMLR Journal 2025 Journal Article

Amphibian: A Meta-Learning Framework for Rehearsal-Free, Fast Online Continual Learning

Gobinda Saha
Kaushik Roy

Online continual learning is challenging as it requires fast adaptation over a stream of data in a non-stationary environment without forgetting the knowledge acquired in the past. To address this challenge, in this paper, we introduce Amphibian - a gradient-based meta-learner that learns to scale the direction of gradient descent to achieve the desired balance between fast learning and continual learning. For this purpose, using only the current batch of data, Amphibian minimizes a meta-objective that encourages alignments of gradients among given data samples along selected basis directions in the gradient space. From this objective, it learns a diagonal scale matrix in each layer that accumulates the history of such gradient alignments. Using these scale matrices Amphibian updates the model online only in the directions having positive cumulative gradient alignments among the data observed so far. With evaluation on standard continual image classification benchmarks, we show that such meta-learned scaled gradient descent in Amphibian achieves better accuracy in online continual learning than relevant baselines while enabling fast learning with less data and few-shot knowledge transfer to new tasks. We also introduce Amphibian-$\beta$ a unified and principled framework for analyzing and understanding the fast learning and continual learning dynamics. Additionally, with loss landscape visualizations, we show such gradient updates incur minimum loss to the old task enabling fast continual learning in Amphibian.

IS Journal 2025 Journal Article

Cognitive Neurosymbolic Artificial Intelligence for Complex Decision-Making: Integrating Foundation Models, Cognitive Architectures, and Knowledge

Yuxin Zi
Kaushik Roy
Amit Sheth

Advances in deep learning and knowledge representation have driven the adoption of intelligent decision support systems in high-stakes domains such as health-care applications. However, traditional statistical approaches often yield opaque, correlation-driven predictions, making it difficult for experts to interpret results and take meaningful actions. To address these limitations, we propose a cognitive neurosymbolic artificial intelligence (AI) framework by combining multimodal perception, neurosymbolic reasoning (with symbolic controls), and knowledge, this paradigm enhances accuracy and robustness in complex decision-making environments. This integrated framework ensures that every recommendation follows human-like decision-making, while being grounded in verifiable, expert knowledge, a crucial attribute for critical application scenarios such as health care. We demonstrate this approach through a mental health support system, showcasing how cognitive neurosymbolic AI can improve decision-making outcomes, personalized interventions, longitudinal tracking, and actionable insights for critical applications.

TMLR Journal 2025 Journal Article

Coresets from Trajectories: Selecting Data via Correlation of Loss Differences

Manish Nagaraj
Deepak Ravikumar
Kaushik Roy

Deep learning models achieve state-of-the-art performance across domains but face scalability challenges in real-time or resource-constrained scenarios. To address this, we propose Correlation of Loss Differences ($\mathtt{CLD}$), a simple and scalable metric for coreset selection that identifies the most impactful training samples by measuring their alignment with the loss trajectories of a held-out validation set. $\mathtt{CLD}$ is highly efficient, requiring only per-sample loss values computed at training checkpoints, and avoiding the costly gradient and curvature computations used in many existing subset selection methods. We develop a general theoretical framework that establishes convergence guarantees for $\mathtt{CLD}$-based coresets, demonstrating that the convergence error is upper-bounded by the alignment of the selected samples and the representativeness of the validation set. On CIFAR-100 and ImageNet-1k, $\mathtt{CLD}$-based coresets typically outperform or closely match state-of-the-art methods across subset sizes, and remain within 1\% of more computationally expensive baselines even when not leading. $\mathtt{CLD}$ transfers effectively across architectures (ResNet, VGG, DenseNet), enabling proxy-to-target selection with $<1\%$ degradation. Moreover, $\mathtt{CLD}$ is stable when using only early checkpoints, incurring negligible accuracy loss. Finally, $\mathtt{CLD}$ exhibits inherent bias reduction via per-class validation alignment, obviating the need for additional stratified sampling. Together, these properties make $\mathtt{CLD}$ a principled, efficient, stable, and transferable tool for scalable dataset optimization.

NeurIPS Conference 2025 Conference Paper

CURE: Concept Unlearning via Orthogonal Representation Editing in Diffusion Models

Shristi Das Biswas
Arani Roy
Kaushik Roy

As Text-to-Image models continue to evolve, so does the risk of generating unsafe, copyrighted, or privacy-violating content. Existing safety interventions - ranging from training data curation and model fine-tuning to inference-time filtering and guidance - often suffer from incomplete concept removal, susceptibility to jail-breaking, computational inefficiency, or collateral damage to unrelated capabilities. In this paper, we introduce CURE, a training-free concept unlearning framework that operates directly in the weight space of pre-trained diffusion models, enabling fast, interpretable, and highly specific suppression of undesired concepts. At the core of our method is the Spectral Eraser, a closed-form, orthogonal projection module that identifies discriminative subspaces using Singular Value Decomposition over token embeddings associated with the concepts to forget and retain. Intuitively, the Spectral Eraser identifies and isolates features unique to the undesired concept while preserving safe attributes. This operator is then applied in a single step update to yield an edited model in which the target concept is effectively unlearned - without retraining, supervision, or iterative optimization. To balance the trade-off between filtering toxicity and preserving unrelated concepts, we further introduce an Expansion Mechanism for spectral regularization which selectively modulates singular vectors based on their relative significance to control the strength of forgetting. All the processes above are in closed-form, guaranteeing extremely efficient erasure in only $2$ seconds. Benchmarking against prior approaches, CURE achieves a more efficient and thorough removal for targeted artistic styles, objects, identities, or explicit content, with minor damage to original generation ability and demonstrates enhanced robustness against red-teaming. Project Page at \url{https: //sites. google. com/view/cure-unlearning/home}.

NeSy Conference 2025 Conference Paper

Learning and Reasoning with Model-Grounded Symbolic Artificial Intelligence Systems

Aniruddha Chattopadhyay
Raj Dandekar
Kaushik Roy

Neurosymbolic artificial intelligence (AI) systems combine neural network and classical symbolic AI mechanisms to exploit the complementary strengths of large-scale, generalizable learning and robust, verifiable reasoning. Numerous classifications of neurosymbolic AI illustrate how these two components can be integrated in distinctly different ways. In this work, we propose reinterpreting instruction-tuned large language models as model-grounded symbolic AI systems—where natural language serves as the symbolic layer, and grounding is achieved through the model’s internal representation space. Within this framework, we investigate and develop novel learning and reasoning approaches that preserve structural similarities to traditional learning and reasoning paradigms. Comprehensive evaluations across complex mathematical reasoning procedures of varying difficulties provide insights into the effectiveness of our approach towards learning efficiency and reasoning reliability.

IS Journal 2025 Journal Article

NeuroSymbolic Knowledge-Grounded Planning and Reasoning in Artificial Intelligence Systems

Amit Sheth
Vedant Khandelwal
Kaushik Roy
Vishal Pallagani
Megha Chakraborty

Decision-support systems in AI-assisted health care require robust, interpretable, and user-centric processes that effectively handle natural language inputs. While large language models (LLMs) excel at generating coherent text, they struggle with complex reasoning and multistep planning tasks. In response, we propose a neurosymbolic framework that integrates LLMs with symbolic knowledge graphs, graph-based reasoners, and constraint-aware planning modules. This hybrid approach leverages LLMs for initial plan formulation while refining outcomes with structured, domain-specific representations that enforce safety standards, ensure regulatory compliance, and maintain logical consistency. Demonstrated through examples in health care and manufacturing, our method bridges the gap between unstructured language generation and formal reasoning, enhancing reliability in high-stakes applications and supporting dynamic, context-aware decision-making. The framework offers a scalable, trustworthy solution for complex, constraint-driven environments. By combining generative creativity with formal logic, our approach addresses the key limitations of LLMs, making it suitable for diverse, high-impact domains.

TMLR Journal 2025 Journal Article

On Inherent Adversarial Robustness of Active Vision Systems

Amitangshu Mukherjee
Timur Ibrayev
Kaushik Roy

Deep Neural Networks (DNNs) are susceptible to adversarial inputs, such as imperceptible noise and naturally occurring challenging samples. This vulnerability likely arises from their passive, one-shot processing approach. In contrast, neuroscience suggests that human vision robustly identifies salient object features by actively switching between multiple fixation points (saccades) and processing surroundings with non-uniform resolution (foveation). This information is processed via two pathways: the dorsal (where) and ventral (what) streams, which identify relevant input portions and discard irrelevant details. Building on this perspective, we outline a deep learning-based active dorsal-ventral vision system and adapt two prior methods, FALcon and GFNet, within this framework to evaluate their robustness. We conduct a comprehensive robustness analysis across three categories: adversarially crafted inputs evaluated under transfer attack scenarios, natural adversarial images, and foreground-distorted images. By learning from focused, downsampled glimpses at multiple distinct fixation points, these active methods significantly enhance the robustness of passive networks, achieving a 2-21 % increase in accuracy. This improvement is demonstrated against state-of-the-art transferable black-box attack. On ImageNet-A, a benchmark for naturally occurring hard samples, we show how distinct predictions from multiple fixation points yield performance gains of 1.5-2 times for both CNN and Transformer based networks. Lastly, we qualitatively demonstrate how an active vision system aligns more closely with human perception for structurally distorted images. This alignment leads to more stable and resilient predictions, with lesser catastrophic mispredictions. In contrast, passive methods, which rely on single-shot learning and inference, often lack the necessary structural understanding.

AAAI Conference 2025 Conference Paper

PIXELS: Progressive Image Xemplar-based Editing with Latent Surgery

Shristi Das Biswas
Matthew Shreve
Xuelu Li
Prateek Singhal
Kaushik Roy

Recent advancements in language-guided diffusion models for image editing are often bottle-necked by cumbersome prompt engineering to precisely articulate desired changes. An intuitive alternative calls on guidance from in-the-wild image exemplars to help users bring their imagined edits to life. Contemporary exemplar-based editing methods shy away from leveraging the rich latent space learnt by pre-existing large text-to-image (TTI) models and fall back on training with curated objective functions to achieve the task. Though somewhat effective, this demands significant computational resources and lacks compatibility with diverse base models and arbitrary exemplar count. On further investigation, we also find that these techniques restrict user control to only applying uniform global changes over the entire edited region. In this paper, we introduce a novel framework for progressive exemplar-driven editing with off-the-shelf diffusion models, dubbed PIXELS, to enable customization by providing granular control over edits, allowing adjustments at the pixel or region level. Our method operates solely during inference to facilitate imitative editing, enabling users to draw inspiration from a dynamic number of reference images, or multimodal prompts, and progressively incorporate all the desired changes without retraining or fine-tuning existing TTI models. This capability of fine-grained control opens up a range of new possibilities, including selective modification of individual objects and specifying gradual spatial changes. We demonstrate that PIXELS delivers high-quality edits efficiently, leading to a notable improvement in quantitative metrics as well as human evaluation. By making high-quality image editing more accessible, PIXELS has the potential to enable professional-grade edits to a wider audience with the ease of using any open-source image generation model.

PDF Details DOI

TMLR Journal 2025 Journal Article

S-TLLR: STDP-inspired Temporal Local Learning Rule for Spiking Neural Networks

Marco Paul E. Apolinario
Kaushik Roy

Spiking Neural Networks (SNNs) are biologically plausible models that have been identified as potentially apt for deploying energy-efficient intelligence at the edge, particularly for sequential learning tasks. However, training of SNNs poses significant challenges due to the necessity for precise temporal and spatial credit assignment. Back-propagation through time (BPTT) algorithm, whilst the most widely used method for addressing these issues, incurs high computational cost due to its temporal dependency. In this work, we propose S-TLLR, a novel three-factor temporal local learning rule inspired by the Spike-Timing Dependent Plasticity (STDP) mechanism, aimed at training deep SNNs on event-based learning tasks. Furthermore, S-TLLR is designed to have low memory and time complexities, which are independent of the number of time steps, rendering it suitable for online learning on low-power edge devices. To demonstrate the scalability of our proposed method, we have conducted extensive evaluations on event-based datasets spanning a wide range of applications, such as image and gesture recognition, audio classification, and optical flow estimation. S-TLLR achieves comparable accuracy to BPTT (within $\pm2\%$ for most tasks), while reducing memory usage by $5-50\times$ and multiply-accumulate (MAC) operations by $1.3-6.6\times$, particularly when updates are restricted to the last few time-steps.

AAAI Conference 2025 Conference Paper

SAP: Corrective Machine Unlearning with Scaled Activation Projection for Label Noise Robustness

Sangamesh Kodge
Deepak Ravikumar
Gobinda Saha
Kaushik Roy

Label corruption, where training samples are mislabeled due to non-expert annotation or adversarial attacks, significantly degrades model performance. Acquiring large, perfectly labeled datasets is costly, and retraining models from scratch is computationally expensive. To address this, we introduce Scaled Activation Projection (SAP), a novel SVD (Singular Value Decomposition)-based corrective machine unlearning algorithm. SAP mitigates label noise by identifying a small subset of trusted samples using cross-entropy loss and projecting model weights onto a clean activation space estimated using SVD on these trusted samples. This process suppresses the noise introduced in activations due to the mislabeled samples. In our experiments, we demonstrate SAP’s effectiveness on synthetic noise with different settings and real-world label noise. SAP applied to the CIFAR dataset with 25% synthetic corruption show upto 6% generalization improvements. Additionally, SAP can improve the generalization over noise robust training approaches on CIFAR dataset by ∼ 3.2% on average. Further, we observe generalization improvements of 2.31% for a Vision Transformer model trained on naturally corrupted Clothing1M.

PDF Details DOI

TMLR Journal 2025 Journal Article

Semantic-Syntactic Discrepancy in Images (SSDI): Learning Meaning and Order of Features from Natural Images

Chun Tao
Timur Ibrayev
Kaushik Roy

Despite considerable progress in image classification tasks, classification models seem unaffected by the images that significantly deviate from those that appear natural to human eyes. Specifically, while human perception can easily identify abnormal appearances or compositions in images, classification models overlook any alterations in the arrangement of object parts as long as they are present in any order, even if unnatural. Hence, this work exposes the vulnerability of having semantic and syntactic discrepancy in images (SSDI) in the form of corruptions that remove or shuffle image patches or present images in the form of puzzles. To address this vulnerability, we propose the concept of "image grammar", comprising "image semantics" and "image syntax". Image semantics pertains to the interpretation of parts or patches within an image, whereas image syntax refers to the arrangement of these parts to form a coherent object. We present a semi-supervised two-stage method for learning the image grammar of visual elements and environments solely from natural images. While the first stage learns the semantic meaning of individual object parts, the second stage learns how their relative arrangement constitutes an entire object. The efficacy of the proposed approach is then demonstrated by achieving SSDI detection rates ranging from 70% to 90% on corruptions generated from CelebA and SUN-RGBD datasets. Code is publicly available at: https://github.com/ChunTao1999/SSDI/.

TMLR Journal 2025 Journal Article

TSkips: Efficiency Through Explicit Temporal Delay Connections in Spiking Neural Networks

Prajna G. Malettira
Shubham Negi
Wachirawit Ponghiran
Kaushik Roy

Spiking Neural Networks (SNNs) with their bio-inspired Leaky Integrate-and-Fire (LIF) neurons inherently capture temporal information. This makes them well-suited for sequential tasks like processing event-based data from Dynamic Vision Sensors (DVS) and event-based speech tasks. Harnessing the temporal capabilities of SNNs requires mitigating vanishing spikes during training, capturing spatio-temporal patterns and enhancing precise spike timing. To address these challenges, we propose _TSkips_, augmenting SNN architectures with forward and backward skip connections that incorporate explicit temporal delays. These connections capture long-term spatio-temporal dependencies and facilitate better spike flow over long sequences. The introduction of _TSkips_ creates a vast search space of possible configurations, encompassing skip positions and time delay values. To efficiently navigate this search space, this work leverages training-free Neural Architecture Search (NAS) to identify optimal network structures and corresponding delays. We demonstrate the effectiveness of our approach on four event-based datasets: DSEC-flow for optical flow estimation, DVS128 Gesture for hand gesture recognition and Spiking Heidelberg Digits (SHD) and Spiking Speech Commands (SSC) for speech recognition. Our method achieves significant improvements across these datasets: up to 18% reduction in Average Endpoint Error (AEE) on DSEC-flow, 8% increase in classification accuracy on DVS128 Gesture, and up to ~8% and ~16% higher classification accuracy on SHD and SSC, respectively.

TMLR Journal 2024 Journal Article

CoDeC: Communication-Efficient Decentralized Continual Learning

Sakshi Choudhary
Sai Aparna Aketi
Gobinda Saha
Kaushik Roy

Training at the edge utilizes continuously evolving data generated at different locations. Privacy concerns prohibit the co-location of this spatially as well as temporally distributed data, deeming it crucial to design training algorithms that enable efficient continual learning over decentralized private data. Decentralized learning allows serverless training with spatially distributed data. A fundamental barrier in such setups is the high bandwidth cost of communicating model updates between agents. Moreover, existing works under this training paradigm are not inherently suitable for learning a temporal sequence of tasks while retaining the previously acquired knowledge. In this work, we propose CoDeC, a novel communication-efficient decentralized continual learning algorithm that addresses these challenges. We mitigate catastrophic forgetting while learning a distributed task sequence by incorporating orthogonal gradient projection within a gossip-based decentralized learning algorithm. Further, CoDeC includes a novel lossless communication compression scheme based on the gradient subspaces. We theoretically analyze the convergence rate for our algorithm and demonstrate through an extensive set of experiments that CoDeC successfully learns distributed continual tasks with minimal forgetting. The proposed compression scheme results in up to 4.8× reduction in communication costs without any loss in performance.

NeurIPS Conference 2024 Conference Paper

Curvature Clues: Decoding Deep Learning Privacy with Input Loss Curvature

Deepak Ravikumar
Efstathia Soufleri
Kaushik Roy

In this paper, we explore the properties of loss curvature with respect to input data in deep neural networks. Curvature of loss with respect to input (termed input loss curvature) is the trace of the Hessian of the loss with respect to the input. We investigate how input loss curvature varies between train and test sets, and its implications for train-test distinguishability. We develop a theoretical framework that derives an upper bound on the train-test distinguishability based on privacy and the size of the training set. This novel insight fuels the development of a new black box membership inference attack utilizing input loss curvature. We validate our theoretical findings through experiments in computer vision classification tasks, demonstrating that input loss curvature surpasses existing methods in membership inference effectiveness. Our analysis highlights how the performance of membership inference attack (MIA) methods varies with the size of the training set, showing that curvature-based MIA outperforms other methods on sufficiently large datasets. This condition is often met by real datasets, as demonstrated by our results on CIFAR10, CIFAR100, and ImageNet. These findings not only advance our understanding of deep neural network behavior but also improve the ability to test privacy-preserving techniques in machine learning.

PDF Details DOI

TMLR Journal 2024 Journal Article

Deep Unlearning: Fast and Efficient Gradient-free Class Forgetting

Sangamesh Kodge
Gobinda Saha
Kaushik Roy

Machine unlearning is a prominent and challenging field, driven by regulatory demands for user data deletion and heightened privacy awareness. Existing approaches involve retraining model or multiple finetuning steps for each deletion request, often constrained by computational limits and restricted data access. In this work, we introduce a novel class unlearning algorithm designed to strategically eliminate specific classes from the learned model. Our algorithm first estimates the Retain and the Forget Spaces using Singular Value Decomposition on the layerwise activations for a small subset of samples from the retain and unlearn classes, respectively. We then compute the shared information between these spaces and remove it from the forget space to isolate class-discriminatory feature space. Finally, we obtain the unlearned model by updating the weights to suppress the class discriminatory features from the activation spaces. We demonstrate our algorithm's efficacy on ImageNet using a Vision Transformer with only ~1.5% drop in retain accuracy compared to the original model while maintaining under 1% accuracy on the unlearned class samples. Furthermore, our algorithm exhibits competitive unlearning performance and resilience against Membership Inference Attacks (MIA). Compared to baselines, it achieves an average accuracy improvement of 1.38% on the ImageNet dataset while requiring up to 10x fewer samples for unlearning. Additionally, under stronger MIA attacks on the CIFAR-100 dataset using a ResNet18 architecture, our approach outperforms the best baseline by 1.8%.

TMLR Journal 2024 Journal Article

DP-ImgSyn: Dataset Alignment for Obfuscated, Differentially Private Image Synthesis

Efstathia Soufleri
Deepak Ravikumar
Kaushik Roy

The availability of abundant data has catalyzed the expansion of deep learning vision algorithms. However, certain vision datasets cannot be publicly released due to privacy reasons. Releasing synthetic images instead of private images is a common approach to overcome this issue. A popular method to generate synthetic images is using Generative Adversarial Networks (GANs) with Differential Privacy (DP) guarantees. However, GAN-generated synthetic images are visually similar to private images. This is a severe limitation, particularly when the private dataset depicts visually sensitive and disturbing content. To address this, we propose a non-generative framework, Differentially Private Image Synthesis (DP-ImgSyn), to generate and release synthetic images for image classification tasks. These synthetic images: (1) have DP guarantees, (2) retain the utility of the private images, i.e., a model trained using synthetic images results in similar accuracy as a model trained on private images, (3) the synthetic images are visually dissimilar to private images. DP-ImgSyn consists of the following steps: First, a teacher model is trained on the private images using a DP training algorithm. Second, public images are used as initialization for the synthetic images which are optimized to align them with the private images. The optimization uses the teacher network's batch normalization layer statistics (mean, standard deviation) to inject information about the private images into the synthetic images. Third, the synthetic images and their soft labels, obtained from the teacher model, are released and can be deployed for neural network training on image classification tasks. Our experiments on various image classification datasets show that when using similar DP training mechanisms, our framework performs better than generative techniques (up to $\approx$ 20% in terms of image classification accuracy).

AAAI Conference 2024 System Paper

GEAR-Up: Generative AI and External Knowledge-Based Retrieval: Upgrading Scholarly Article Searches for Systematic Reviews

Kaushik Roy
Vedant Khandelwal
Valerie Vera
Harshul Surana
Heather Heckman
Amit Sheth

This paper addresses the time-intensive nature of systematic reviews (SRs) and proposes a solution leveraging advancements in Generative AI (e.g., ChatGPT) and external knowledge augmentation (e.g., Retrieval-Augmented Generation). The proposed system, GEAR-Up, automates query development and translation in SRs, enhancing efficiency by enriching user queries with context from language models and knowledge graphs. Collaborating with librarians, qualitative evaluations demonstrate improved reproducibility and search strategy quality. Access the demo at https://youtu.be/zMdP56GJ9mU.

PDF Details DOI

TMLR Journal 2024 Journal Article

Homogenizing Non-IID Datasets via In-Distribution Knowledge Distillation for Decentralized Learning

Deepak Ravikumar
Gobinda Saha
Sai Aparna Aketi
Kaushik Roy

Decentralized learning enables serverless training of deep neural networks (DNNs) in a distributed manner on multiple nodes. One of the key challenges with decentralized learning is heterogeneity in the data distribution across the nodes. Data heterogeneity results in slow and unstable global convergence and therefore poor generalization performance. In this paper, we propose In-Distribution Knowledge Distillation (IDKD) to address the challenge of heterogeneous data distribution. The goal of IDKD is to homogenize the data distribution across the nodes. While such data homogenization can be achieved by exchanging data among the nodes sacrificing privacy, IDKD achieves the same objective using a common public dataset across nodes without breaking the privacy constraint. This public dataset is different from the training dataset and is used to distill the knowledge from each node and communicate it to its neighbors through the generated labels. With traditional knowledge distillation, the generalization of the distilled model is reduced due to misalignment between the private and public data distribution. Thus, we introduce an Out-of-Distribution (OoD) detector at each node to label a subset of the public dataset that maps close to the local training data distribution. Our experiments on multiple image classification datasets and graph topologies show that the proposed IDKD scheme is more effective than traditional knowledge distillation and achieves state-of-the-art generalization performance on heterogeneously distributed data with minimal communication overhead.

IS Journal 2024 Journal Article

Neurosymbolic AI for Enhancing Instructability in Generative AI

Amit Sheth
Vishal Pallagani
Kaushik Roy

Generative AI, especially via large language models (LLMs), has transformed content creation across text, images, and music, showcasing capabilities in following instructions through prompting, largely facilitated by instruction tuning. Instruction tuning is a supervised fine-tuning method where LLMs are trained on datasets formatted with specific tasks and corresponding instructions. This method systematically enhances the model’s ability to comprehend and execute the provided directives. Despite these advancements, LLMs still face challenges in consistently interpreting complex, multistep instructions and generalizing them to novel tasks, which are essential for broader applicability in real-world scenarios. This article explores why neurosymbolic AI offers a better path to enhance the instructability of LLMs. We explore the use of a symbolic task planner to decompose high-level instructions into structured tasks, a neural semantic parser to ground these tasks into executable actions, and a neuro-symbolic executor to implement these actions while dynamically maintaining an explicit representation of state.

IS Journal 2024 Journal Article

Neurosymbolic Value-Inspired Artificial Intelligence (Why, What, and How)

Amit Sheth
Kaushik Roy

The rapid progression of artificial intelligence (AI) systems, facilitated by the advent of large language models (LLMs), has resulted in their widespread application to provide human assistance across diverse industries. This trend has sparked significant discourse centered around the ever-increasing need for LLM-based AI systems to function among humans as a part of human society. Toward this end, neurosymbolic AI systems are attractive because of their potential to enable and interpretable interfaces for facilitating value-based decision making by leveraging explicit representations of shared values. In this article, we introduce substantial extensions to Kahneman’s System 1 and System 2 framework and propose a neurosymbolic computational framework called value-inspired AI (VAI). It outlines the crucial components essential for the robust and practical implementation of VAI systems, representing and integrating various dimensions of human values. Finally, we further offer insights into the current progress made in this direction and outline potential future directions for the field.

NeurIPS Conference 2024 Conference Paper

OPEL: Optimal Transport Guided ProcedurE Learning

Sayeed Shafayet Chowdhury
Soumyadeep Chandra
Kaushik Roy

Procedure learning refers to the task of identifying the key-steps and determining their logical order, given several videos of the same task. For both third-person and first-person (egocentric) videos, state-of-the-art (SOTA) methods aim at finding correspondences across videos in time to accomplish procedure learning. However, to establish temporal relationships within the sequences, these methods often rely on frame-to-frame mapping, or assume monotonic alignment of video pairs, leading to sub-optimal results. To this end, we propose to treat the video frames as samples from an unknown distribution, enabling us to frame their distance calculation as an optimal transport (OT) problem. Notably, the OT-based formulation allows us to relax the previously mentioned assumptions. To further improve performance, we enhance the OT formulation by introducing two regularization terms. The first, inverse difference moment regularization, promotes transportation between instances that are homogeneous in the embedding space as well as being temporally closer. The second, regularization based on the KL-divergence with an exponentially decaying prior smooths the alignment while enforcing conformity to the optimality (alignment obtained from vanilla OT optimization) and temporal priors. The resultant optimal transport guided procedure learning framework (`OPEL') significantly outperforms the SOTA on benchmark datasets. Specifically, we achieve 22. 4\% (IoU) and 26. 9\% (F1) average improvement compared to the current SOTA on large scale egocentric benchmark, EgoProceL. Furthermore, for the third person benchmarks (ProCeL and CrossTask), the proposed approach obtains 46. 2\% (F1) average enhancement over SOTA.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Continual Learning with Scaled Gradient Projection

Gobinda Saha
Kaushik Roy

In neural networks, continual learning results in gradient interference among sequential tasks, leading to catastrophic forgetting of old tasks while learning new ones. This issue is addressed in recent methods by storing the important gradient spaces for old tasks and updating the model orthogonally during new tasks. However, such restrictive orthogonal gradient updates hamper the learning capability of the new tasks resulting in sub-optimal performance. To improve new learning while minimizing forgetting, in this paper we propose a Scaled Gradient Projection (SGP) method, where we combine the orthogonal gradient projections with scaled gradient steps along the important gradient spaces for the past tasks. The degree of gradient scaling along these spaces depends on the importance of the bases spanning them. We propose an efficient method for computing and accumulating importance of these bases using the singular value decomposition of the input representations for each task. We conduct extensive experiments ranging from continual image classification to reinforcement learning tasks and report better performance with less training overhead than the state-of-the-art approaches.

PDF Details DOI

AAAI Conference 2023 System Paper

Demo Alleviate: Demonstrating Artificial Intelligence Enabled Virtual Assistance for Telehealth: The Mental Health Case

Kaushik Roy
Vedant Khandelwal
Raxit Goswami
Nathan Dolbir
Jinendra Malekar
Amit Sheth

After the pandemic, artificial intelligence (AI) powered support for mental health care has become increasingly important. The breadth and complexity of significant challenges required to provide adequate care involve: (a) Personalized patient understanding, (b) Safety-constrained and medically validated chatbot patient interactions, and (c) Support for continued feedback-based refinements in design using chatbot-patient interactions. We propose Alleviate, a chatbot designed to assist patients suffering from mental health challenges with personalized care and assist clinicians with understanding their patients better. Alleviate draws from an array of publicly available clinically valid mental-health texts and databases, allowing Alleviate to make medically sound and informed decisions. In addition, Alleviate's modular design and explainable decision-making lends itself to robust and continued feedback-based refinements to its design. In this paper, we explain the different modules of Alleviate and submit a short video demonstrating Alleviate's capabilities to help patients and clinicians understand each other better to facilitate optimal care strategies.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Global Update Tracking: A Decentralized Learning Algorithm for Heterogeneous Data

Sai Aparna Aketi
Abolfazl Hashemi
Kaushik Roy

Decentralized learning enables the training of deep learning models over large distributed datasets generated at different locations, without the need for a central server. However, in practical scenarios, the data distribution across these devices can be significantly different, leading to a degradation in model performance. In this paper, we focus on designing a decentralized learning algorithm that is less susceptible to variations in data distribution across devices. We propose Global Update Tracking (GUT), a novel tracking-based method that aims to mitigate the impact of heterogeneous data in decentralized learning without introducing any communication overhead. We demonstrate the effectiveness of the proposed technique through an exhaustive set of experiments on various Computer Vision datasets (CIFAR-10, CIFAR-100, Fashion MNIST, and ImageNette), model architectures, and network topologies. Our experiments show that the proposed method achieves state-of-the-art performance for decentralized learning on heterogeneous data via a 1-6% improvement in test accuracy compared to other existing techniques.

TMLR Journal 2023 Journal Article

Neighborhood Gradient Mean: An Efficient Decentralized Learning Method for Non-IID Data

Sai Aparna Aketi
Sangamesh Kodge
Kaushik Roy

Decentralized learning algorithms enable the training of deep learning models over large distributed datasets, without the need for a central server. The current state-of-the-art decentralized algorithms mostly assume the data distributions to be Independent and Identically Distributed (IID). In practical scenarios, the distributed datasets can have significantly different data distributions across the agents. This paper focuses on improving decentralized learning on non-IID data with minimal compute and memory overheads. We propose Neighborhood Gradient Mean (NGM), a novel decentralized learning algorithm that modifies the local gradients of each agent using self- and cross-gradient information. In particular, the proposed method averages the local gradients with model-variant or data-variant cross-gradients based on the communication budget. Model-variant cross-gradients are derivatives of the received neighbors’ model parameters with respect to the local dataset. Data-variant cross-gradient derivatives of the local model with respect to its neighbors’ datasets. The data-variant cross-gradients are aggregated through an additional communication round. We theoretically analyze the convergence characteristics of NGM and demonstrate its efficiency on non-IID data sampled from various vision and language datasets. Our experiments demonstrate that the proposed method either remains competitive or outperforms (by 0-6%) the existing state-of-the-art (SoTA) decentralized learning algorithm on non-IID data with significantly less compute and memory requirements. Further, we show that the model-variant cross-gradient information available locally at each agent can improve the performance on non-IID data by 3-20% without additional communication costs.

IS Journal 2023 Journal Article

Neurosymbolic Artificial Intelligence (Why, What, and How)

Amit Sheth
Kaushik Roy
Manas Gaur

Humans interact with the environment using a combination of perception—transforming sensory inputs from their environment into symbols, and cognition—mapping symbols to knowledge about the environment for supporting abstraction, reasoning by analogy, and long-term planning. Human perception-inspired machine perception, in the context of artificial intelligence (AI), refers to large-scale pattern recognition from raw data using neural networks trained using self-supervised learning objectives such as next-word prediction or object recognition. On the other hand, machine cognition encompasses more complex computations, such as using knowledge of the environment to guide reasoning, analogy, and long-term planning. Humans can also control and explain their cognitive functions. This seems to require the retention of symbolic mappings from perception outputs to knowledge about their environment. For example, humans can follow and explain the guidelines and safety constraints driving their decision making in safety-critical applications such as health care, criminal justice, and autonomous driving.

AAAI Conference 2022 Conference Paper

Oscillatory Fourier Neural Network: A Compact and Efficient Architecture for Sequential Processing

Bing Han
Cheng Wang
Kaushik Roy

Tremendous progress has been made in sequential processing with the recent advances in recurrent neural networks. However, recurrent architectures face the challenge of exploding/vanishing gradients during training, and require significant computational resources to execute back-propagation through time. Moreover, large models are typically needed for executing complex sequential tasks. To address these challenges, we propose a novel neuron model that has cosine activation with a time varying component for sequential processing. The proposed neuron provides an efficient building block for projecting sequential inputs into spectral domain, which helps to retain long-term dependencies with minimal extra model parameters and computation. A new type of recurrent network architecture, named Oscillatory Fourier Neural Network, based on the proposed neuron is presented and applied to various types of sequential tasks. We demonstrate that recurrent neural network with the proposed neuron model is mathematically equivalent to a simplified form of discrete Fourier transform applied onto periodical activation. In particular, the computationally intensive back-propagation through time in training is eliminated, leading to faster training while achieving the state of the art inference accuracy in a diverse group of sequential tasks. For instance, applying the proposed model to sentiment analysis on IMDB review dataset reaches 89. 4% test accuracy within 5 epochs, accompanied by over 35x reduction in the model size compared to LSTM. The proposed novel RNN architecture is well poised for intelligent sequential processing in resource constrained hardware.

AAAI Conference 2022 Conference Paper

Spiking Neural Networks with Improved Inherent Recurrence Dynamics for Sequential Learning

Wachirawit Ponghiran
Kaushik Roy

Spiking neural networks (SNNs) with leaky integrate and fire (LIF) neurons, can be operated in an event-driven manner and have internal states to retain information over time, providing opportunities for energy-efficient neuromorphic computing, especially on edge devices. Note, however, many representative works on SNNs do not fully demonstrate the usefulness of their inherent recurrence (membrane potential retaining information about the past) for sequential learning. Most of the works train SNNs to recognize static images by artificially expanded input representation in time through rate coding. We show that SNNs can be trained for practical sequential tasks by proposing modifications to a network of LIF neurons that enable internal states to learn long sequences and make their inherent recurrence resilient to the vanishing gradient problem. We then develop a training scheme to train the proposed SNNs with improved inherent recurrence dynamics. Our training scheme allows spiking neurons to produce multi-bit outputs (as opposed to binary spikes) which help mitigate the mismatch between a derivative of spiking neurons’ activation function and a surrogate derivative used to overcome spiking neurons’ non-differentiability. Our experimental results indicate that the proposed SNN architecture on TIMIT and LibriSpeech 100h speech recognition dataset yields accuracy comparable to that of LSTMs (within 1. 10% and 0. 36%, respectively), but with 2× fewer parameters than LSTMs. The sparse SNN outputs also lead to 10. 13× and 11. 14× savings in multiplication operations compared to GRUs, which are generally considered as a lightweight alternative to LSTMs, on TIMIT and LibriSpeech 100h datasets, respectively.

EAAI Journal 2011 Journal Article

Towards nonideal iris recognition based on level set method, genetic algorithms and adaptive asymmetrical SVMs

Kaushik Roy
Prabir Bhattacharya
Ching Y. Suen

We present algorithms for iris segmentation, feature extraction and selection, and iris pattern matching. To segment the inner boundary from a nonideal iris image, we apply a level set based curve evolution approach using the edge stopping function, and to detect the outer boundary, we employ the curve evolution approach using the regularized Mumford–Shah segmentation model with an energy minimization algorithm. Daubechies wavelet transform (DBWT) is used to extract the textural features, and genetic algorithms (GAs) are deployed to select the subset of informative features by combining the valuable outcomes from the multiple feature selection criteria without compromising the recognition accuracy. To speed up the matching process and to control the misclassification error, we apply a combined approach called the adaptive asymmetrical support vector machines (AASVMs). The parameter values of SVMs are also optimized in order to improve the overall generalization performance. The verification and identification performance of the proposed scheme is validated using the UBIRIS Version 2, the ICE 2005, and the WVU datasets.